Taxonomy of Three Oxytrichids (Protozoa, Ciliophora, Hypotrichia), With Establishment of the New Species Rubrioxytricha guangzhouensis spec. nov.

Two oxytrichid ciliates collected from China, Oxytricha lithofera Foissner, 2016 and Rubrioxytricha haematoplasma (Blatterer and Foissner, 1990) Berger, 1999 were investigated based on living observations, protargol preparations, and molecular analyses. The Chinese population of O. lithofera shares highly distinctive features with the type population, e.g., lithosomes, abutting macronuclear nodules, straight undulating membranes, and long dorsal bristles. The morphology and morphogenesis of our new isolate of R. haematoplasma corresponds well with the type population. The 18S rRNA gene sequences of both species were obtained, and the molecular phylogeny of the genera Oxytricha Bory de Saint-Vincent in Lamouroux et al., 1824 and Rubrioxytricha Berger, 1999 was analyzed. In addition, we found that the Guangzhou population of R. haematoplasma described by Chen et al. (2015) can be distinguished from the type population and our isolate by the reddish-brown cortical granules (vs. lemon yellowish to greenish) and mitochondria-like granules present (vs. absent), as well as the slightly red cytoplasm (vs. slightly orange) and the marine or brackish water habitat (vs. freshwater). In addition, a 20-bp divergence in their 18S rRNA gene sequences indicates that they are not conspecific. Thus, we establish a new species, Rubrioxytricha guangzhouensis, for R. haematoplasma sensu Chen et al. (2015).


INTRODUCTION
The family Oxytrichidae Ehrenberg, 1830, including more than 200 valid species, is one of the most morphologically and morphogenetically diverse hypotrichid ciliate groups (Berger, 1999(Berger, , 2008Foissner et al., 2002;Shao et al., 2011;Foissner, 2016;Hu et al., 2019;Dong et al., 2020;Wang et al., 2020a). The genus Oxytricha Bory de Saint-Vincent in Lamouroux et al., 1824, the type and time-honored genus of the family Oxytrichidae, comprises almost 40 nominal species isolated from various habitats worldwide and has been a melting pot for flexible 18cirri hypotrichs with caudal cirri for over one century (Berger, 1999;Shao et al., 2011Shao et al., , 2014Weisse et al., 2013;Foissner, 2016;. Even though many studies have been carried out to delimit the genus more clearly (Berger, 1999;Shao et al., 2011Shao et al., , 2015Foissner, 2016), it is still a polyphyletic genus with species irregularly distributed across the molecular tree (Schmidt et al., 2007;Paiva et al., 2009;Shao et al., 2014). The incongruence between morphology and molecular information suggests that this taxon is a genus with high genetic divergence (Berger, 1999;Foissner, 2016). Oxytricha lithofera, a highly distinctive species with a combination of features absent in any other described congeners (s. l.), was originally described by Foissner (2016) based on a population collected from moss on stones and soil from an old coral cave in the Golfete de Cuare del Indio, Venezuela. Here, we provide not only a description of the first Chinese population of O. lithofera but also the first analysis of the 18S rRNA gene sequence for this species.
In a revision of oxytrichids, Berger (1999) transferred two Oxytricha species, viz., Oxytricha haematoplasma Blatterer and Foissner, 1990 and Oxytricha ferruginea Stein, 1859, to the newly established genus Rubrioxytricha Berger, 1999, with the former as the type species. Rubrioxytricha was mainly characterized by the homogeneously colored cytoplasm (slightly orange to reddish in R. haematoplasma and rusty brown in R. ferruginea), the presence of one or two caudal cirri, and four to five dorsal kineties. Three more Rubrioxytricha species have since been reported, namely R. indica Naqvi et al., 2006, R. tsinlingensis Chen et al., 2017, and R. guamensis Kumar et al., 2018(Naqvi et al., 2006Chen et al., 2017;Kumar et al., 2018). Recently, Chen et al. (2015) reported a brackish/marine population collected from Guangzhou, China, under the name R. haematoplasma. Herein, we describe a freshwater population of R. haematoplasma based on a detailed investigation of morphology and morphogenesis. A comprehensive comparison between R. haematoplasma populations as well as congeners is provided, and a new species is established. Updated phylogenetic analyses based on 18S rRNA gene sequences were also performed, and the non-monophyly of the genera Oxytricha and Rubrioxytricha are discussed.

Sample Collection, Observation, and Terminology
A brackish water sample containing Oxytricha lithofera was collected from a sandy beach near Zhuhai Fishing Girl Statue (22 • 15 41 N; 113 • 35 16 E; Figures 1A,B) in Zhuhai, southern China, on 26 May 2014, when the water temperature was about 21 • C and the salinity was about 3 . The sample was maintained at about 22 • C in Petri dishes with rice grains to enrich bacteria as a food source for the ciliates. Oxytricha lithofera was identified in this sample on 6 June 2014, when the salinity of the sample reached 6 , and the population collapsed over the next 3 days. Investigations for O. lithofera were made with specimens from raw cultures rather than from cloned individuals. However, O. lithofera is easily distinguished by its long dorsal bristles and no other Oxytricha-like morphotypes were present in the protargol preparations, indicating that our morphological and molecular studies deal with the same species.
Rubrioxytricha haematoplasma was collected on 8 and 11 April 2016 from a freshwater lake in Wuhan Botanical Garden, Chinese Academy of Sciences (30 • 32 57 N; 114 • 25 51 E; Figures 1A,C), Wuhan, China, when the water temperature was about 21 • C. Sponge cubes, with the length about 10 cm, submerged at a depth of about 1 m for about 2 weeks, were used as artificial substrates to allow colonization by ciliates. Samples were collected according to Luo et al. (2018). Cells were selected and cultivated in Petri dishes at about 22 • C with Volvic mineral water and rice grains. Although attempts to establish clonal cultures were unsuccessful, no other hypotrichid morphotypes were present in the protargol preparations, indicating that our morphological, morphogenetic, and molecular studies deal with the same species.
Live cells were isolated with micropipettes and then observed and measured with bright field and differential interference contrast microscopy at 40-1,000 × magnification using a Zeiss Axioplan 2 imaging microscope equipped with a Zeiss AxioCam HRc camera. The protargol staining method of Wilbert (1975) was used to reveal the nuclear apparatus and the ciliature. Counts, measurements, and photographs for protargol-stained specimens were made at a magnification of 1,000 × using an Olympus BX53 compound microscope (Tokyo, Japan) equipped with a Pixelink M5DC-SE-CYL camera (Ottawa, Canada). Drawings were made at a magnification of 1,000 × or 1,250 × with the aid of a drawing attachment or photomicrographs as templates. In the drawings of the morphogenetic stages, parental structures are shown as outlines, whereas new ones are shaded in black. Terminology is according to Berger (1999) and Foissner (2016).
The statistical probability of the monophyletic hypothesis of the genus Rubrioxytricha was evaluated using AU tests (Shimodaira, 2002) in the CONSEL package (Shimodaira and Hasegawa, 2001).

Deposition of Voucher Specimens
Two voucher slides with protargol-stained specimens were deposited in the Laboratory of Protozoology, OUC, China, with registration numbers LXT2014052601-01∼02.
Morphology Based on the Zhuhai Population (Figures 2A-P, 3A-J and Table 1) Cells 80-115 × 25-40 µm in vivo, ellipsoid in outline with both ends rounded, left margin more or less convex, right margin usually more or less straight (Figures 2A,B   Adoral zone extending 35-51% of body length (average 42%) in protargol preparations, composed of 21-29 membranelles ( Table 1), with cilia 13-15 µm long in vivo. Paroral and endoral membrane almost equal in length, more or less straight, optically side by side, or superimposed, with paroral extending slightly more anteriorly than endoral (Figures 2E,L,N).
Ventral cirral pattern as shown in Figures 2E,K,L,N. Three frontal cirri with the right one located behind the distal end of the adoral zone, easily misinterpreted as an adoral membranelle. One buccal cirrus, located anterior to the endoral membrane and right of the anterior end of the paroral membrane. Usually four frontoventral cirri (only one out of 20 cells observed with three, Figure 2K) arranged in V shape, three of them (IV/3, VI/3, VI/4) arranged in a line, cirrus III/2 slightly above the level of cirrus  VI/3. Three postoral ventral cirri arranged in a triangular pattern. Two pretransverse ventral cirri, with the posterior one almost at the same level as transverse cirrus III/1. Five enlarged transverse cirri, arranged in a check-mark shape, with cilia 18-20 µm long in vivo, conspicuously protruding beyond the posterior end of the cell (Figures 2A, 3A-E). Single marginal row on either body side, separated by a distinct gap, left one with 14-18 cirri, right one with 13-16 cirri, with cilia about 13 µm in length.
Invariably six dorsal kineties (DK): DK1, the leftmost kinety, strongly shortened anteriorly, reaches the posterior end, with an indeed very indistinct gap in most specimens (Figures 2F,G, double arrowheads), only six out of 25 cells observed with a slightly conspicuous gap (Figures 2O,P, arrowhead); DK2 almost extending the entire body length; DK3 extends to the anterior body end and splits to form a short row of DK4 (on an average of five dikinetids) at the subterminal end, usually one or two dikinetids between DK3 and DK4; DK5 and DK6, the dorsomarginal kineties, shortened anteriorly, with DK5 reaching the mid-body, DK6 extends slightly above the midbody (Figures 2F,G,M). Bristles conspicuous in vivo, even at low magnification (Figures 3A-C,J, arrows), anterior bristles of DK1-3 about 5 µm long, gradually increase to 9 µm posteriorly; bristles of DK4 usually 8-9 µm long, bristles of DK5 and DK6 usually 5-7 µm long. Three long caudal cirri, with cilia 18-20 µm long, one each at end of DK1, 2, and 4, optically positioned in gap between ends of marginal rows (Figures 2F,G

,M).
Rubrioxytricha haematoplasma (Blatterer andFoissner, 1990) Berger, 1999 Deposition of Voucher Specimens Voucher slides with protargol-stained specimens were deposited in the Laboratory of Protozoology, OUC, China, with registration numbers LXT2016040802-01∼03 and LXT2016041102-01∼02. (Figures 4A-N and Table 1) Cells 125-160 × 40-60 µm in vivo. Body soft, flexible and not contractile. Body long ellipsoid in shape, both anterior and posterior ends rounded, some with broad ends, some with narrower ends, right and left margins more or less convex (Figures 4A,D,G-I). Dorsoventrally flattened about 2:1. Single contractile vacuole slightly behind the buccal vertex and ahead of the mid-body, about 45% of body length, at the left cell margin, approximately 12 µm in diameter in diastole, two longitudinally oriented collecting canals present (Figures 4A,D,G-I). Cortical granules of single type, spherical, 0.8-1.0 µm across, lemon yellowish to greenish, arranged in lines or small groups then in long lines on both body sides (Figures 4B,C,J-L, slightly squeezed cells). Cytoplasm slightly orange (Figures 4J,L), usually packed with numerous irregularly shaped granules (2-6 µm-sized) and food vacuoles, containing diatoms, sometimes rendering cells yellow-brown to slightly darkish at low magnification. Consistently two macronuclear nodules, positioned in the mid-body and more or less left of midline, with the anterior one almost at the level of the buccal vertex, each nodule elongate ellipsoidal in shape, on average 25 × 10 µm after protargol preparation. Usually two ellipsoidal micronuclei, 5-8 × 3-5 µm in size, with one attached to or near each macronuclear nodule (Figures 4F,M,N). Locomotion by slowly crawling on the substrate, occasionally swimming with cells rotating around the longitudinal axis.
Ventral cirral pattern as shown in Figures 4E,M. Three slightly enlarged frontal cirri continuous with frontoventral cirri arranged in a V-shaped pattern, rightmost three (IV/3, VI/3, VI/4) almost arranged in a line, cirrus III/2 slightly ahead level of the posteriormost one (cirrus IV/3) and distant from cirrus VI/3. Single buccal cirrus, almost at the same level with the right frontal cirrus. Three postoral ventral cirri and two pretransverse ventral cirri. On average five slightly enlarged transverse cirri, with cilia 20-22 µm long, extending slightly behind the rear end of the cell, arranged in a check-mark shape, just behind the pretransverse ventral cirri. Single left and right marginal row, more or less confluent at the posterior end.

Morphogenesis (Figures 5A-H, 6A-H and Table 1)
In a very early divider (Figures 5A, 6A), stomatogenesis commences with appearance of the oral primordium as groups of basal bodies between the buccal vortex and transverse cirri, including a large group just left to the postoral ventral cirri. At this stage, all parental cirri remain intact and do not contribute to the formation of the oral primordium. Somewhat later (Figures 5B, 6B), the anterior two postoral ventral cirri (IV/2, V/4) dedifferentiate and contribute to the oral primordium. In an early divider (Figures 5C, 6C), the posterior postoral ventral cirrus (V/3) is no longer recognizable and the oral primordium gives rise to new adoral membranelles. Simultaneously, the parental undulating membranes dedifferentiate anteriorly to form the undulating membranes anlage (UMA) (Figure 5C, arrow) and the frontal-ventral-transverse cirral anlagen (FVTA) begin to form. In a middle divider (Figures 5E, 6D), two sets of FVTA are formed. Subsequently (Figures 5G, 6F), the UMA forms the new undulating membranes and gives rise to the left frontal cirrus. At the same time, the other streaks differentiate into new cirri in the pattern of 3:3:3:4:4 from left to right. The parental adoral zone, which is inherited by the proter, remains unchanged throughout the whole process (Figures 5C,E,G). The marginal anlagen develop by dedifferentiation of the parental structures, and the right marginal anlage of the proter appears earlier than others (Figures 5C,E, 6C,D). Then, all the anlagen extend longitudinally and generate new cirri in both proter and opisthe (Figures 5G, 6F).
Dorsal morphogenesis occurs by two groups of primordia: one group develops by dedifferentiation of the parental structures at the middle stage, that is, each of dorsal kineties 1-3 forms an anlage in both proter and opisthe (Figures 5F, 6E, arrows). Later, all the anlagen lengthen at both ends, and one caudal cirrus is generated at the end of dorsal kinety 3 (Figures 5H, 6H). The second group, that is, the dorsomarginal kinety primordia, develops de novo to the anterior right of the right marginal row anlage during the late stage (Figures 5G, 6F, arrowheads). During the late stage of the ontogenetic process, the macronuclear nodules fuse with one another and the micronuclei fuse into a single mass as well (Figures 5H, 6G).

18S rRNA Gene Sequences and Phylogenetic Analyses
The 18S rRNA gene sequence of Oxytricha lithofera (excluding both primers) has been deposited in GenBank with accession number MT364897. It has a length of 1,726 bp and a G + C content of 45.48%. The closest sequence to O. lithofera is Urosoma salmastra (KF951419), with the sequence similarity of 98.3% (28 nucleotide difference). The 18S rRNA gene sequence of Rubrioxytricha haematoplasma (excluding both primers) has been deposited in GenBank with accession number MT364898. It has a length of 1,724 bp and a G + C content of 45.01%. The 18S rRNA gene sequences of all six Rubrioxytricha isolates differ from each other by 0 to 26 nucleotides, with sequence identities from 98.4 to 100% ( Table 2).
In the phylogenetic trees, Oxytricha species are separated into several clades. Oxytricha lithofera is sister to U. salmastra (KF951419) and distant from the type species of the genus, O. granulifera (AF508762), and O. granulifera chiapasensis (KX889988) although this relationship is not robust as indicated by the low support in the ML tree (ML/BI, 54%/1.00). All species of Rubrioxytricha fall within a wellsupported assemblage (ML/BI, 91%/1.00) although this genus

Identification of the Chinese Population of Oxytricha lithofera and Emended Diagnosis of the Species
Based on a population collected from Golfete de Cuare del Indio (salinity 10 ), Venezuela, Foissner (2016) reported a highly distinctive species, O. lithofera, with a combination of features not present in any other Oxytricha species. Our population, collected from a sandy beach (salinity 3-6 ) of Zhuhai, shares most of the following typical features with the Venezuelan population: (1) cirral pattern; (2) straight and parallel undulating membranes; (3) abutting macronuclear nodules, usually with a globular micronucleus attached to each; (4) large lithosomes present; (5) dorsal bristles increasing in length from the anterior end to the posterior end of dorsal kineties, up to 10 µm long; and (6) mitochondria-like granules present. Compared with the type population, the Zhuhai population has (1) a relatively larger body size in vivo (80-115 × 25-40 vs. 60-85 × 23-32 µm) and after protargol impregnation (60-105 × 29-51 vs. 53-75 × 20-28 µm) (shrinkage differed between the two methods); (2) more adoral membranelles (21-29 vs. 19-21); and (3) more marginal cirri (left, 14-18 vs. 12-15; right, 13-16 vs. 9-12). However, all of these differences, most of which overlap with each other, can be considered as population-dependent geographic variations (Berger, 1999;Chen et al., 2020). One difference that cannot be ignored is that there is a one-bristle-wide gap in dorsal kinety 1 of the type population, while the gap in most of the specimens of the Zhuhai population is absent. This may be seen as an intraspecific or intrapopulational difference. Therefore, we consider the Zhuhai population conspecific with the Venezuelan population. Based on the information of the Zhuhai population and the type population, an emended diagnosis is supplied here.

Emended Diagnosis of Oxytricha lithofera
Size

Identification of the Populations of Rubrioxytricha haematoplasma
Hitherto, five Rubrioxytricha species have been reported, namely R. haematoplasma, R. ferruginea, R. indica, R. tsinlingensis, and R. guamensis (Berger, 1999;Naqvi et al., 2006;Chen et al., 2017;Kumar et al., 2018). The type species of the genus, R. haematoplasma, was originally reported by Blatterer and Foissner (1990) based on a German population. Subsequently, Shin and Kim (1993) described a Korean population. Recently, Chen et al. (2015) reported a brackish/marine population collected from Guangzhou, China, based on detailed morphological, morphogenetic, and phylogenetic analyses. The main features of the Wuhan population correspond well with those of the type population of R. haematoplasma, e.g., body shape, cirral pattern, dorsal ciliature, especially the color of cortical granules and cytoplasm. There is a tiny difference in terms of the distribution pattern of cortical granules, that is, the cortical granules of the type population are arranged in regular lines as shown in the drawings (Figure 5d in Blatterer and Foissner, 1990), while the cortical granules in the Wuhan population seem more or less irregular, with some in small groups and then in long lines. In addition, the postoral ventral cirri in the type population are distributed in a line, while those in the Wuhan population are in a V pattern.
However, we consider that these differences are not enough to separate them as different species. The Wuhan population of R. haematoplasma can be easily distinguished from the other four congeners by the characteristics shown in Table 3, while the differences of the 18S rRNA gene sequences of the five Rubrioxytricha members ranges from 1 to 20 nucleotides ( Table 2). It is noteworthy that there is only one nucleotide difference of the 18S rRNA gene between R. haematoplasma and R. tsinlingensis, while morphologically, they have distinct different dorsal ciliature (four dorsal kineties with only one caudal cirrus vs. six dorsal kineties with three caudal cirri) ( Table 3). The highly conserved nature of the 18S rRNA gene has been noted in other oxytrichid morphospecies as well (Kumar et al., 2017;Fan et al., 2021). Thus, the Wuhan population can be identified as R. haematoplasma.
There was no information about cortical granules or cytoplasm color of the Korean population (Shin and Kim, 1993). The differences in numbers of adoral membranelles and in the dorsal ciliature were considered as geographical variations by Berger (1999) and population differences by Chen et al. (2015). Herein, we tentatively accept these opinions until more information (cortical granules, cytoplasm, molecular data) is available for the Korean population. Chen et al. (2015) did mention that the paroral and endoral membranes of the Guangzhou population of R. haematoplasma are more curved than those of the type population. However, this might be a preparation artifact, because photomicrographs of live specimens do not show a wide and deep buccal cavity as discussed by Kumar et al. (2018), and the undulating membranes of the early dividers (Figures 3a,b in Chen et al., 2015) do not show a strong degree of curvature either. Therefore, there is no distinct difference of the undulating membranes among the populations, all of which are arranged in an Australocirrus pattern. Compared with the Guangzhou population described by Chen et al. (2015), the Wuhan population and the type population of R. haematoplasma show significant differences in the following aspects: (1) cortical granules (lemon yellowish to greenish in color vs. reddish-brown), (2) mitochondria-like granules (absent vs. present), (3) color of the cytoplasm (slightly orange vs. slightly red), and (4) the habitat (freshwater vs. brackish/marine) ( Table 3). Hence, we consider the Guangzhou population as a different species from R. haematoplasma. Furthermore, the 18S rRNA gene sequences (20 nucleotide difference between the Wuhan and the Guangzhou populations, Table 2) indicate that these two populations represent separate species.
In terms of cytoplasm color, cortical granules, cirral pattern, and dorsal ciliature, the Guangzhou population of R. haematoplasma should be compared with the other four species of the genus. The Guangzhou population of R. haematoplasma can be distinguished from all the others by the color of cortical granules (reddish-brown vs. R. ferruginea, brownish; R. guamensis, yellowish; R. indica, dark green; R. tsinlingensis, yellow-green) and the presence of mitochondrialike granules (vs. absent in all the congeners). For more differences among all the species, see Table 3. Thus, the Guangzhou population represents a separate species of the genus Rubrioxytricha.

Remarks
As discussed above, a new species should be established for the Guangzhou isolate which was described under the name R. haematoplasma by Chen et al. (2015). Chen et al. (2015) stated "the Chinese population came from two shrimp-farming ponds, one brackish (salinity 8.6 ) and the other marine (34.4 ). However, the brackish water was created artificially by mixing together marine and freshwater, and material was regularly transferred between the two ponds, so it is unclear whether the Chinese population of R. haematoplasma was originally from a marine or a freshwater habitat." We think it is possible that the Guangzhou isolate was originally from the marine pond and then transferred to the brackish pond. Therefore, we tentatively treat the marine pond as the type locality.

Etymology
The species-group name guangzhouensis refers to the geographic locality where the ciliate population was discovered.

Type Specimens
The protargol-stained slide containing the holotype specimen (Figures 7C-E) and 10 paratype slides were deposited in the Laboratory of Protozoology, OUC, China, with registration numbers CXM08112701-01∼11.
In the present study, we supply somewhat more detailed morphogenetic information for R. haematoplasma based on the Wuhan population. Chen et al. (2015) provided a detailed morphogenetic description for R. guangzhouensis spec. nov. Naqvi et al. (2006) reported the morphogenetic process of R. indica and Kumar et al. (2018) provided brief notes on the ontogenesis of R. guamensis. All the abovementioned populations share the following morphogenetic features: (1) the parental adoral zone of membranelles is completely inherited, unchanged, by the proter; (2) the undulating membranes anlage of the proter originates by dedifferentiation of the parental structures; (3) parental buccal cirrus (II/2), two frontoventral cirri (III/2, IV/3), and three postoral ventral cirri (IV/2, V/3, V/4) contribute to the formation of the anlagen; (4) the frontal-ventral-transverse (FVT) cirral anlagen are probably formed via a primary pattern and the FVT cirri are formed in a typical Oxytricha pattern, i.e., in the pattern of 1:3:3:3:4:4 from left to right; (5) marginal anlagen develop intrakinetally within the parental structures; (6) dorsal kineties are formed in the Urosomoida pattern, i.e., dorsomarginal row present and without fragmentation of kinety 3; (7) a single caudal cirrus (seldom two caudal cirri in R. guamensis) originates at the end of dorsal kinety 3; and (8) two macronuclear nodules fuse into a single mass during the morphogenetic process. This combination of features reveals a highly conservative morphogenetic pattern of the above Rubrioxytricha species (Naqvi et al., 2006;Chen et al., 2015;Song and Shao, 2017;Kumar et al., 2018).

Phylogenetic Analyses
As described in previous studies, Oxytricha and Urosoma species are irregularly distributed over the phylogenetic trees, revealing that both the genus Oxytricha and Urosoma are nonmonophyletic and lack obvious synapomorphies (Schmidt et al., 2007;Paiva et al., 2009;Shao et al., 2014Shao et al., , 2019Gao et al., 2016;Jung and Berger, 2019;Kaur et al., 2019;Dong et al., 2020;Wang et al., 2020b;Xu et al., 2020). In the present phylogenetic trees (Figure 8), O. lithofera is placed sister to U. salmastra (KF951419). Morphologically, they do have the same features as follows: flexible body, single marginal row on each body side, three caudal cirri, two macronuclear nodules, and two micronuclei, which, however, are shared by most of the oxytrichids (Berger, 1999;Wang et al., 2017). Moreover, O. lithofera can be easily distinguished from U. salmastra by an adoral zone of membranelles more or less in Stylonychia pattern (vs. typical Gonostomum pattern), having more transverse cirri (five vs. four), and more dorsal kineties (six, kinety fragmentation probably present vs. four, kinety fragmentation absent), which all have been considered as genus-level diagnostic features for the oxytrichids (Berger, 1999;Shao et al., 2015;Wang et al., 2020b). In addition, the grouping of O. lithofera and U. salmastra is far from stable as this relationship does not receive high support in the ML analysis. The topology may very likely be changed when additional molecular data from more taxa are available. All species of Rubrioxytricha are intermingled with P. monilata (KT192639) and P. flava (KP266627) in a highly supported assemblage, indicating that Rubrioxytricha is paraphyletic. The hypothesis that Rubrioxytricha species form a monophyletic group was rejected by the AU test, although the internal relationships are not yet robustly resolved. The close relationship between Rubrioxytricha and Polystichothrix-Pseudogastrostyla is morphologically supported as they all have a flexible body, one marginal row on each body side, and brightly colored cortical granules. However, Rubrioxytricha differs significantly from Polystichothrix and Pseudogastrostyla in having a typical oxytrichid pattern of 18 FVT cirri (vs. more than 18 FVT cirri in short lines) (Fan et al., 2015;Luo et al., 2017). The phylogenetic placement of the genera Polystichothrix and Pseudogastrostyla might be explained by the assumption that those taxa with more than 18 FVT cirri evolved by a secondary increase in the numbers of FVT anlagen and/or additional cirri originated from some of the anlagen (Berger, 1999;Luo et al., 2017). Moreover, the dorsal ciliature pattern, which also carries an informative phylogenetic signal, like the ventral cirral pattern (Berger, 1999;Shao et al., 2011Shao et al., , 2015Foissner, 2016), shows great variety in the Rubrioxytricha-Polystichothrix-Pseudogastrostyla clade: (1) four dorsal kineties in R. guamensis, R. haematoplasma, R. indica, R. guangzhouensis spec. nov., and Pseudogastrostyla, with one dorsomarginal kinety present, dorsal kinety fragmentation absent, only one caudal cirrus (seldom two caudal cirri in R. guamensis) present at the end of dorsal kinety 3; (2) five dorsal kineties in R. ferruginea and Polystichothrix, very likely including one dorsomarginal kinety and dorsal kinety 4 originated by dorsal kinety 3 fragmentation, one or two caudal cirri present in R. ferruginea and caudal cirri lacking in Polystichothrix; (3) six dorsal kineties in R. tsinlingensis, probably with two dorsomarginal rows included and dorsal kinety 4 originated by dorsal kinety 3 fragmentation, three caudal cirri in total, one each at the posterior end of dorsal kineties 1, 2, and 4 (Berger, 1999;Naqvi et al., 2006;Chen et al., 2015Chen et al., , 2017Fan et al., 2015;Luo et al., 2017;Kumar et al., 2018).
The incongruence between morphology and molecular data suggests that Oxytricha and Rubrioxytricha are genera with high genetic divergence, which need to be broken up into multiple new genera. At present, there are over 1,000 nominal hypotrich ciliates ; however, the 18S rRNA gene sequences are only available for about 30% of this group (NCBI database). Thus, molecular data from additional taxa and even phylogenomic data are needed to reveal the natural phylogeny of this group of ciliates.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are publicly available. This data can be found here: GenBank (https: //www.ncbi.nlm.nih.gov/genbank/) with accession numbers MT364897-MT364898.

AUTHOR CONTRIBUTIONS
XL and JH collected the samples. XL carried out the experiments (live observation, protargol impregnation, and DNA amplification) and analyses (morphometrics and phylogenetic analyses). XC and WS supervised the work. All authors wrote the manuscript and approved the final version.